Function of the HspA protein

3.2.1. Expression of hspA in E. coli

Using the Qiagen expression system (Qiaexpresss kit), hspA gene was first cloned into the expression vector pQE9 (pSH11) and transformed into E. coli strain M15. However, most of the fusion protein formed was insoluble and was found in inclusion bodies, which were detected by immunoelectron microscopy (cooperation with H. Lünsdorf, GBF, Germany, data not shown). This fusion protein was purified under denaturing conditions. Efforts were made to get soluble fusion protein by removal of urea from the protein preparation by stepwise dialysis (urea concentration in the dialysis buffer is reduced from 6 M to 0 M, gradually).

However, the fusion protein precipitated during dialysis when the concentration of urea was lower than 3 M (data not shown). To get a soluble fusion protein preparation, a strain from an expression system of Invitrogen (ThioFusion^TM Expression kit) was employed and the hspA

carried by pSH11 was expressed in E. coli strain GI698 at room temperature. More than half of the fusion protein formed under this condition was soluble. The possible reason for the improved solubility is the slow induction of hspA expression at lower temperature. It does not exclude that the host strain GI698 has a genetic background that somehow facilitates the fusion protein to fold correctly. However, the manufacturer did not offer such information (personal communication with M. Younessian, Invitrogen). The soluble fusion protein was purified under native conditions. Theoretically, such soluble fusion protein should be folded in a correct way and thus can be used for biochemical assays.

3.2.2. The quaternary structure of HspA_His

HspA_His has an additional His-tag with 6 Histidine residues and an enterokinase recognition site at the N-terminus of the HspA protein, yielding an extra sequence encoding about 2 kDa polypeptide added to HspA. Enterokinase (Boehringer Mannheim) digestion was performed to remove the His-tag. However, HspA_His itself was degraded during enterokinase treatment (data not shown).

The oligomeric structure is characteristic for small heat shock proteins. The oligomer is the functional unit. The mechanism by which small heat shock proteins interact with unfolded proteins and thus prevent their aggregation is unknown. Leroux R. M. (Leroux et al., 1997b) established a model to explain the structure-function relationship of small heat shock proteins using Hsp16.2 from C. elegans as a model system (Fig. 3.2).

Fig. 3.2. Model for sHSP oligomeric structure and interaction with unfolded polypeptides (Leroux et al., 1997b). a, the proposed central cavity;

b, N-terminal domain interactions; c, the interaction site between sHSP and unfolded protein; d, whether C-terminal extension is involved in the function of sHSP is unknown;

e, how the subunits of sHSP assemble cooperatively into an oligomer is unknown.

According to this model, the N-terminal domain of sHSPs is buried in the central cavity as a consequence of the assembly of small heat shock proteins. The Hsp16.2 with poly His–tag was compared with the recombinant wild-type Hsp16.2. Both proteins assembled into

large oligomeric complexes although the size of the oligomers were somewhat different. The molecular mass of the Hsp16.2 wild-type oligomer is 550 kDa, while the size of Hsp16.2His is 680 kDa, as judged by SEC. In addition, the Hsp16.2_His exhibited the same chaperone activities as the wild-type Hsp16.2 in vitro, suggesting that the N-terminus of Hsp16.2 can accommodate at least an additional 4-kDa of heterologous sequence per subunit without affecting its chaperone properties. Native His-tagged Hsp16.2 could not bind to Ni-agarose affinity resin.

In the case of HspAHis, the His-tagged HspA formed also a large oligomeric complex, which can interact with unfolded CS. However, in contrast to Hsp16.2_His, native HspAHis

binds to Ni-agarose affinity resin. This means that the assembly of HspA_His does not match this model. How the N-terminal domain is arranged in the oligomeric complex of small heat shock proteins is far from clear. As mentioned in the Introduction, the N-terminal 32 residues are disordered in the complex of MjHSP16.5.

3.2.3. The chaperone activities of HspAHis

HspAHis interacts with chemically unfolded CS to prevent its aggregation in the solution.

Bovine α-carystallin and mouse Hsp25 do not prevent chemically unfolded CS aggregation.

In contrast to HspAHis,_α-crystallin and Hsp25 suppress the aggregation of the unfolded insulin B-chain. In the course of insulin B-chain unfolding, partial structured intermediates are present. At the beginning of CS refolding process, the chemically denatured CS is completely unfolded. This suggests that at least some small heat shock proteins interact selectively with certain structures of the substrate. Hsp16.2 from C. elegans can interact with both thermally denatured and chemically denatured CS. It has reduced selectivity for the substrate structure.

3.2.4. The interaction of HspAHis with unfolded CS

HspAHis prevents the reactivation of unfolded CS, suggesting that a stable complex is formed between HspAHis and unfolded CS. Such complexes may be productive or dead-end intermediates.

The chaperone function of small heat shock proteins may be cooperative with other chaperones as shown in the case of murine Hsp25 (Ehrnsperger et al., 1997) and E. coli IbpB (Veinger et al., 1998). When the complex formed by Hsp25 and thermally denatured CS was supplemented with Hsp70 and ATP, about 15% of CS were reactivated. Heat denatured malate dehydrogenase (MDH) was released from IbpB-MDH complex after supplementation

of DnaK/DnaJ/GrpE (KJE) chaperones and ATP. About 10% reactivation of MDH was observed. Moreover, GroEL/GroES (LS) chaperonines accelerated the rate of KJE-mediated refolding of IbpB-released MDH. Refolding of urea-heat denatured lactate dehydrogenase displayed a similar dependence on the IbpB/KJE/LS chaperone network.

In the case of HspAHis, Hsp70 and ATP did not cause the release of bound CS from the complex of HspAHis with unfolded CS. The cofactor for HspA_His is unknown.

3.2.5. The physiological function of HspA

To elucidate the physiological function of HspA, a hspA deletion mutant was constructed.

Four hspA deletion mutant strains were obtained. This small amount is due to the lower recombination efficiency when using linear DNA for transformation by electroporation. Three of the mutants showed a wild-type S. aurantiaca phenotype. One of the mutants showed abnormal rippling and an altered fruiting body (data not shown). M. Heidelbach described similar observations in his Inaugural-dissertation (ZMBH, Germany, 1992). Three strains with inactivated hspA were obtained in his work; one showed wild-type phenotype, the other two were unable to form fruiting bodies. Obviously, the alternation of the development of the hspA mutants is not due to the inactivation or deletion of hspA but due to unknown mutations since various phenotypes were detected. Furthermore, anti-HspA-sera did not affect the fruiting body formation of DW4/3-1 wild-type cells (data not shown).

As pointed out in the Introduction, small heat shock proteins from different organism have evolved different biological activities (Arrigo and Landry, 1994). The major small heat shock protein of Saccharomyces cerevisiae is Hsp26. Its expression is induced by heat shock, starvation or high concentration of salt. Another small heat shock protein identified in yeast cells is Hsp42 that shares high homology with Hsp26. In contrast to Hsp26, Hsp42 expression is more dependent on an increased salt concentration and on starvation. It is expressed also in unstressed cells. However, neither disruption of the gene encoding HSP26 nor a hsp26/hsp42 double mutant had a detectable phenotype, even under stress conditions (Petko and Lindquist, 1986; Wotton et al., 1996).

The physiological role of HspA in S. aurantiaca is still enigmatic and the hspA deletion mutant did not offer any clue. However, the in vitro chaperone assay proved that HspA_His has chaperone properties, suggesting HspA may have also a protective function in vivo.

Immunoelectron microscopy revealed HspA to be mainly distributed along the cytoplasmic membrane or the cell wall of the stressed cells. But, it was impossible to localise

HspA differentially in the outer membrane or at the inner side or outside of the cytoplasmic membrane by immunoelectron microscopy.

Periplasm is an important compartment of gram-negative bacteria that participates in cell physiological functions, e.g. solute transport and protein secretion. It has been observed that some periplasmic substrate-binding proteins have chaperone like properties. In addition to their role in transport and chemotaxis, they might help protein folding and renaturation in the periplasm (Richarme and Caldas, 1997).

During vegetative growth and development, myxobacteria secrete numerous extracellular proteins including proteinases and developmental signals (Dworkin, 1996). For example, the C signal has been demonstrated to be localised in the extracellular matrix by immunogold electron microscopy. Possibly, it is associated with the extracellular fibrils. The C signal as well as the spore coat protein U have a signal peptide for secretion (Gollop et al., 1991;

Shimkets and Rafiee, 1990), while another spore coat protein S has no signal peptide.

Myxobacteria have two extracellular appendages: pili and fibrils. Other proteins involved in gliding and cell-cell interaction should be expressed on the cell surface. It is possible that HspA plays a role in the stabilisation of extracellular or periplasmatic proteins under heat shock or in protein folding in the course of development.